Figure 6. Alternative PD models from different animal species show similar metabolic dysregulations in serum.

(A) Score plot of the OPLS-DA model built with ¹HNMR spectra of serum samples from α-synuclein rats versus first predictive and first orthogonal components (α-synuclein OPLS model). Three groups are clearly discriminated, mimicking 3 different stages of PD (sham operated, n = 12; prodomal-like, n = 11 clinical-like n = 11). R2Y = 0.929; Q2 = 0.518; 1 predictive and 3 orthogonal components; CV ANOVA, P = 0.008. (B) Relative amplitude of most discriminating metabolites in the α-synuclein OPLS model, i.e., acetoacetate, betaine, BHB, creatine, glycine, myoinositol (myo), pyruvate,and serine in sham-operated (n = 12), prodromal-like (n = 11), and clinical-like (n = 11) animals. (C) Score plot of the OPLS-DA model built with ¹HNMR spectra of serum samples from nonhuman MPTP primates versus first predictive and first orthogonal components (MPTP OPLS-DA model). Sham-operated (n = 6) and clinical-like (n = 8) animals are clearly discriminated. R2Y = 0.998; Q² = 0.963; 1 predictive and 1 orthogonal component; CV ANOVA P = 0.0005. (D) Relative amplitude of most discriminating metabolites in the MPTP OPLS-DA model, i.e., acetoacetate, alanine, betaine, BHB, creatine, lactate, pyruvate, and valine in sham-operated (n = 6) and MPTP (n = 8) animals. Data are represented as mean ± SEM, mixed model followed by Tukey’s post hoc test or Mann-Whitney test. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001.